The invention discloses a traffic speed prediction method based on an improved state space model, and the method comprises the steps: 1), inputting an adjacent matrix and historical traffic speed data into a self-adaptive graph structure + space-time fusion feature layer, and enabling the self-adaptive graph structure + space-time fusion feature layer to learn a more realistic adjacent matrix on the basis of an original adjacent matrix, the adjacent matrix and historical traffic speed data are fused, and an embedding layer of fusion time and space features is obtained; and 2) inputting the Embedding into the Backbone, obtaining future traffic speed characteristics by the Backbone according to the state matrix, the input matrix, the output matrix and the discretized state-space equation, and obtaining a predicted value of the future traffic speed according to the future traffic speed characteristics through a multilayer MLP network of a decoder. According to the method, the adaptive graph structure is adopted, more appropriate and real space information can be learned, and therefore a better prediction effect is achieved.
本发明公开了一种基于改进状态空间模型的交通速度预测方法,包括:1)将邻接矩阵和历史交通速度数据输入自适应图结构+时空融合特征层,自适应图结构+时空融合特征层会在原先邻接矩阵的基础上学习一个更贴切真实的邻接矩阵,该邻接矩阵和历史交通速度数据进行融合,获得融合时间空间特征的嵌入层Embedding;2)将Embedding输入进Backbone,Backbone根据状态矩阵、输入矩阵、输出矩阵以及离散化的状态空间方程获得未来交通速度特征,未来交通速度特征通过解码器的多层MLP网络得到未来交通速度的预测值。本发明采用了自适应图结构,可以学习到更贴切真实的空间信息,从而达到更好的预测效果。
Traffic speed prediction method based on improved state space model
基于改进状态空间模型的交通速度预测方法
2023-07-04
Patent
Elektronische Ressource
Chinesisch
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen |
Traffic flow prediction method based on improved space-time Transform
Europäisches Patentamt | 2022
|Traffic speed prediction method based on space-time dependency enhancement network
Europäisches Patentamt | 2024
|Time-space traffic state prediction method based on Transform network
Europäisches Patentamt | 2021
|Traffic speed prediction method based on multi-spatial scale space-time Transform
Europäisches Patentamt | 2023
|